Analysing In-Nozzle Flows Using Partial Nozzle Models: Martin Volmajer, MSC Breda Kegl, PH.D
Analysing In-Nozzle Flows Using Partial Nozzle Models: Martin Volmajer, MSC Breda Kegl, PH.D
Analysing In-Nozzle Flows Using Partial Nozzle Models: Martin Volmajer, MSC Breda Kegl, PH.D
Abstract
The geometry of the diesel fuel injection nozzle and fuel flow characteristics in the nozzle significantly
affects the processes of fuel atomisation, combustion and formation of pollutant emissions in a diesel
engine. To improve the processes of fuel injection and spray formation CFD packages are commonly
used. Since CPU times are often high, partial models are used for analysis. The results of the analysis
by using partial models are model dependent, so the differences between the results using different
models can occur. In this paper CFD analysis for different partial models of a 4-hole nozzle were
made. The objective of the research was to find the proper model to be used for fast analysis of
existing nozzle geometry. Several partial models of the nozzle were made. The results were compared
with the results of measurements at steady state conditions.
1 INTRODUCTION
The fuel injection system of the diesel engine plays the dominant role on the fuel spray
formation, which affects the combustion and pollutant formation processes. It is well known
that injector nozzle flow has a significant role on the spray, but it still not very well
investigated. The major problem represents the dimension of the nozzle channel flow areas,
which have a size of less then one millimetre and pressures which can exceed even 200 MPa.
Under this conditions measurements or other observations of the flow in the nozzle are very
difficult. The measurements are more or less limited to the measurements of the nozzle flow
coefficients at the steady state conditions [1-3] or observations and measurements of the flow
in the scaled-up transparent models [1][4-7].
On the other side computational fluid dynamics (CFD) packages gained on importance in
recent years, since by using the three-dimensional analysis many parameters can be calculated
at every single cell of the modelled nozzle. The main problem that occurs at CFD analysis is
the required computational time. To shorten the time, needed for calculation of a single
analysis, simplified models representing a cut of 1/2, 1/4, etc of the nozzle (for ex. see [1][2])
are used. Results obtained from these models as the matter of fact represent the state in this
simplified model, which can differ from the actual conditions in the real size models.
This is why the main goal of research presented in this paper was to find the proper simplified
model of some existing four-hole nozzle that will be used for fast analysis of internal
geometry changes. Fuel flow analyses were made at several simplified models and the results
of the CFD analyses were compared to the results of the flow coefficient at the steady state
conditions.
2 THEORETICAL BACKGROUNDS
Analyses were made for four-hole nozzle with a sac volume and sharp edges at the nozzle
hole inlet sides. The dimensions of the tested nozzle are presented in Table 1 and Figure 1.
Table 1: Nozzle dimensions
Nozzle hole diameter dd 0,375 mm
Nozzle hole channel length ls 1,0 mm
Sac chamber diameter DE 1,0 mm
Needle seat diameter DA 1,1 mm
Nozzle hole inclination angle (holes #1, #4) α1,4 49°
Nozzle hole inclination angle (holes #2, #3) α2,3 95°
Needle tip cone angle α 95°
Needle seat cone angle σ 60°
Angle of the radial hole distribution between #1,#2 and #3,#4 δ12=δ34 76°
Angle of the radial hole distribution between #2,#3 and #4,#1 δ23=δ41 104°
Maximal needle lift hmax 0,35 mm
The flow coefficient, nevertheless its simplicity, represents one of the most important values,
related to the fuel injection conditions at the nozzle. It is defined as the ratio between the
measured or real (Vreal) and theoretical (Vth) volume flow injected through the nozzle.
According to the Bernoulli equation, the theoretical outflow velocity can be derived from the
pressure difference (∆p) and fuel density (ρ):
V Vreal
µ = real = . (eq.1)
Vth 2 ⋅ ∆p
Ad ⋅
ρ
Ad represents the sum of the nozzle hole cross-section area, while ρ is the density of the fluid.
Following the presented equations, higher flow coefficient values result in a larger quantity of
fuel injected per time unit, higher outflow velocities and better fuel spray atomisation.
3 ANALYSIS
3.1 Experimental apparatus
The fuel flow coefficient-measuring device is made at the University's Engine Research
Laboratory. The testing device (for detail presentation see [3]) is used for measuring the fuel
flow coefficients at the steady state conditions. The testing device consist of the low and high
pressure pump, each driven by an electric motor, pressure chamber for reduction of pressure
waves, pressure regulation valve, three way electromagnet valve for changing the direction of
the flow into the measuring Plexiglas valve or directly into the fuel tank. The nozzle i.e.
nozzle holder is fixed in such a way that it enables measuring of flow coefficients for different
nozzles. The needle lift is fixed by using a micrometer. The time of the measurement is set by
a timer, which controls the opening and closing of the electromagnetic valve. The volume of
the fluid injected at predefined time is measured at constant pressure drop for different needle
lifts. The measured flow coefficient is derived form equation 1, where the real volume flow is
calculated from measured volume and duration of the measurement according to equation 2:
V
Vreal = measured . (eq.2)
tmeasurement
The calculations of the measured value of flow coefficient with corresponding uncertainties
were calculated for every needle lift from ten repeated measurements of volume of the fluid,
pressure drop and temperature.
To analyse the flow characteristics of the in-nozzle flow several different nozzle models,
representing nozzle lifts from 0.05 to 0.35 mm, were made. Since some analysis [8][9] shown,
that the pressure drop in nozzle is significant only in the area of the needle seat, sac chamber
and nozzle holes, the meshes were modelled only for the above-mentioned parts. For the
maximal needle lift of 0,35 mm, four different nozzle models, representing real size, one half
and one quarter (two models, one for each hole) of the nozzle were made. Since the first
analyses have shown no significant differences between the results of the real size and one
half model and because of the nozzle symmetry, for other needle lifts only half- and quarter
size models were made (See Figure 2 and Figure 3). For needle lifts of 0,35 and 0,05 mm also
one-eighth nozzle models, each representing one half of the hole with symmetry boundary
condition at the cross section plane, were made.
According to the steady state analysis conditions, pressure boundary conditions at the in- and
outlet are specified. The fluid used for analysis is the calibration fluid according to ISO
4113[10], with a temperature of 313 K, the density 825 kg/m3 and kinematic viscosity of 2.6
mm2/s. K-ε turbulence model is employed. Since maximal velocities are much smaller than
the speed of sound, the fluid is supposed to be uncompressible.
4 RESULTS
4.1 Results of the numerical analysis
The velocity profiles and pressure distributions derived from the CFD analyses for different
nozzle models are presented on following figures.
velocity pressure
hole #3 hole #3
hole #1 hole #1
velocity pressure
hole #2 hole #2
hole #4 hole #4
Figure 2: Real size nozzle model results: Velocity flow field and pressure distribution
The velocity profiles and pressure distributions in nozzle holes with equal inclination angle
are almost identical. The results indicate an already known fact that the outflow velocities are
higher at the holes with smaller inclination angles, which results in a higher flow coefficient
at those holes. At the nozzle holes #2 and #3 there is the region of low pressure at the upper
nozzle hole inlet area, which indicates that flow in the nozzle probably cavitates already at the
pressures that are much lower that the actual injection pressure.
As already stated the results of the analysis on the real size and one-half model of the nozzle
showed no significant differences. Velocity fields, velocity profiles on the outlet and pressure
distributions are comparable. The calculated values of the flow coefficients are 0,685 for real
size and 0,686 for one half model, where the cpu times for one half model are only 50 % of
the cpu needed for the real size model calculation.
velocity pressure
hole #3
hole #3
hole #1
hole #1
Figure 3: One-half nozzle model results at h=0,35 mm: Velocity flow field and pressure distribution
h=0.3 mm h=0.35 mm
velocity
0 m/s 150
h=0.05 mm h=0.2 mm
h=0.1 mm
h=0.3 mm h=0.35 mm
pressure
0 Pa 11e7
velocity pressure
u [m/s]
150.00
100.00
50.00
0.00
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Point
model 1/1 model 1/2 model 1/4 model 1/8
x [mm]
0.225 0.225
x [mm]
1-2 model
0.15 0.15
1-8 model
1-1 model
0.075 0.075
0 0
150 100 50 0 0 50 100 150
Figure 8: Comparison of the calculated velocities (m/s) profiles at the central axis of the outflow (hole 2#-
left, hole 1#-right) at maximum needle lift (h=0.35 mm)
The calculated values of the flow coefficients are presented in table 2. As it was mentioned
above the nozzle holes with smaller inclination angle yield higher flow coefficients. So the
value of flow coefficient for a single hole (at 1/4 and 1/8) with smaller inclination angle is
higher then for two holes at 1/2 model and vice versa for the holes with higher inclination
angle.
Table 2: Comparison of the calculated flow coefficients values for different models
Model 1/8 1/4 1/2 1/1
Holes 1# 2# 1# 2# 1# & 2# all
h=0.05 mm 0.136 0.146 0.147 0.143 0.146
h=0.1 mm 0.447 0.417 0.425
h=0.2 mm 0.681 0.623 0.647
h=0.3 mm 0.72 0.664 0.685
h=0.35 mm 0.766 0.680 0.715 0.672 0.686 0,685
The results of the experiment are in detail presented in [3]. In this paper only the flow
coefficient results are presented in Figure 8.
0.8
0.7 1/1
0.6 1/2
0.5 1/4-1#
0.4 1/4-3#
µ
0.3 1/8-1#
0.2 1/8-3#
Meas.
0.1
Emp.
0
0 0.1 0.2 0.3
h [mm]
5 CONCLUSIONS
The presented results have shown that the one-half model meets the results of the analysis on
the real size model and gives good comparison with the results of the measurement. The same
can be stated also for the results of the one-quarter model, which can be with no restrictions
used for fast analyses in case of geometry changes, since the flow characteristics are rather
similar to those of the real size model. The only negative fact is that the calculated flow
coefficient couldn't be compared directly to the measurement of the total flow coefficient.
The results of one-eighth model differ slightly from the others. Consequently the
adequateness of this model could not be confirmed clearly. Nevertheless we believe those
models can be also used for some very quick analyses.
6 REFERENCES
[1] T. Yoda, T. Tsuda, Inflence of Injection Nozzle Improvement on DI Diesel Engine, SAE
paper 970356
[2] Y.Oishi et.al., A computational Study into the Effect of the Injection Nozzle Inclintion
Angle on the Flow Characteristics in Nozzle Holes, SAE paper 920580
[3] M.Volmajer, B.Kegl, Experimental and numerical analysis of fuel flow in diesel engine
injection nozzle, J. of. KONES Internal combustion engines, vol. 8, no. 1/2, p. 9-16
[4] Arcoumanis C. et al., Investigation of Cavitation in a Vertical Multi-Hole Injector, SP-
1415, p.181-198, SAE, Warrendale, 1999
[5] C.Arcoumanis et. al., Analysis of the Flow in the Nozzle of a Vertical Multi Hole Diesel
Engine Injector, SAE paper 980811
[6] M.Kato et al., Flow Analysis in Nozzle Hole in Consideration of Cavitation, SAE
970052
[7] K.Melcher, I.Komaroff, Experimentelle Utersuchung der Stroemung durch
Dieseleinspritzduesen in stationaer betriebenen Grossmodel, Bosch Techn. Berichte 5
(1976) 4
[8] Y.Oishi et.al., A Computational Study into the Effects of the Injection Nozzle Inclination
Angle on the Flow Characteristics in the Nozzle Holes, SAE paper 920580
[9] K.Melcher, J.Chomiak, Experimentelle Untersuchung der Stroemung durch
Dieseleinspritzduesen in stationaer betriebenen Grossmodel, Bosch Techn.Berichte
5(1976)4
[10] ISO-4113:1988, Road Vehicles- Calibration Fluid Diesel Injection Equipment
[11] H.Hardenberg, Die Nadelhubabhaengigkeit der Durchflussbeiwerte von Lochduessen
fuer Direkteinspritzdieselmotoren, MTZ 46(1985) 4, p. 143-146